一种基于相似度的混合型多属性决策方法

Technique of hybrid multiple attribute decision making based on similarity degree to ideal solution

研究了实数、区间数、语言值和不确定语言值相结合的混合型多属性决策问题。基于逼近理想解的排序方法(TOPSIS方法)确定理想方案,并定义了一种新的相似度,将各类型属性值的相似度计算规范到统一的度量空间,并且指出该相似度的几何意义,说明用该方法描述方案相似程度的合理性。将待选方案与理想方案比较,进行排序择优。最后,给出两个算例,证明了该算法的有效性。

A hybrid multiple attribute decision making problem with real number, interval number, linguistic and uncertain linguistic value is analysed. Based on TOPSIS method, alternatives are compared with ideal solution by defining a new type of similarity degree which regulates all types of attribute value into a metric space uniformly and whose geometrical meaning is presented to explain the rationality of similarity degree. Then the order of alternatives is sorted to select the best one. Finally, two examples are provided to proue the validity of this method.